基于霍夫变换和相位相关的图像配准方法

本文结合霍夫变换和相位相关技术来处理图像配准问题.根据1D相位相关技术,提出了一种通过搜索角度脊来得到旋转参数的方法;将待处理图像去旋转后,再采用2D相位相关来得到平移参数.本文方法可以充分利用FFT 的高效运算.实验结果表明此方法能准确有效地处理具有旋转和平移的图像配准问题.

Image Registration Based on Hough Transform and Phase Correlation

In this paper, Hough transform and phase correlation was combined to solve the problem of image registration. Based on 1D phase correlation, a method was proposed to obtain the angle of rotation by searching an angle ridge. After de-rotation, the traditional 2D phase correlation was imported to get the translation. The method can take full advantage of the high efficiency of FFT. Experiments show that the proposed method can detect the rotation and translation between images efficiently and exactly.